ballistafandomcom-20200215-history
User blog:Samuli.seppanen/Fixes, bugs and performance improvements
After a brief pause and many small fixes I got back to shooting the cheiroballistra. Several things were fixed between this and the previous shooting session. The single most important fix was making a small notch for the field-frame bar at the end of the little ladder bar. This notch was designed to prevent the field-frames from moving due to the torque exerted upon them during cocking: These notches worked fairly well, but need to be made deeper; apparently after ~25 shots both front field-frame bars had slipped from the notches: Probably part of the reason for the slippage is the semi-roundness of one side of the field-frame bars. The other reason is that the nothing - except the two wooden wedges to some extent - is keeping the tenons against the field-frame bars. Small pieces of hardwood pushed between the tenons and the pi-brackets should fix this issue for good. The little arch stayed in place with no issues. This is not very surprising, given that it is under much lesser stress compared to the little ladder. That is, unless the top of the field-frames are made rigid using some sort of struts. Or that's how my theory goes, at least. Another important fix was adding a crude auto-locking mechanism to the handle. Essentially the only change was to make the retaining pins point backwards instead of forward. This means that when the handle contacts the pin, it will rise until it finally drops behind the pin. These pictures illustrate the mechanism perfectly: This mechanism works in fully automated fashion, but only if the slider is fairly close to horizontal. Unfortunately this is not the case during cocking, so a small tap with hand is necessary to lock it, but the mechanism is still much better than previously, when the handle had to be operated entirely manually. Thanks to my friend Boris for pointing out this very simple improvement! The claw and slider attachment was also improved: Now the bowstring will rest on a slightly elevated platform, and is much less likely to slip under the claw's fingers. This allows use of thinner and lighter bowstring with much less fear of dry-fires. The claw's fingers still rise a little as power is increased, but so far that hasn't been a problem. One of the dovetailed portions of the case was glued again, as it had become partially loose after the ballista fell to the ground. As original surface preparation of the dovetailed part - and the core of the case - had been suboptimal, I added further safety by screwing the dovetailed pieces to the case with 5 screws each: The arms were also upgraded. Their weight is now 70.92 grams, as opposed to 93.13 grams. Even more importantly most of their mass is now much closer to the spring bundle than previously. The bars are made of ~3.5 mm rectangular, forged reinforcement steel, and they weigh 18.27 grams. The old 5 mm bars weighed 48.32 grams. Unfortunately the new arms are a bit of patchwork, because I ended up redesigning them without having the ability to remake them properly. Here's a picture of the new arms before the hoop is attached: And here's a picture of the hoop and a temporary T-shaped shield protecting the cone and the bar: The essential change is the use of a open hoop made from fairly thick (5 mm) rectangular steel instead of a thinner, riveted hoop like previously. The open hoop actually works beautifully, because it can be tightened trivially in a vise or using heavy pliers, provided the hoop is of fairly soft steel which has a low yield point. So far the hoop has not slipped at all, even though it's just squeezed tight, not sunk to any grooves in the cone. However, as can be seen from the above picture the spring cords still try their best to slip over it. I will probably have to tie the half-bundles with twine to discourage this kind of behavior. If this does not solve the problem, I can make the ends of the (new) hoops extend a bit further away from the belly of the cone. In the final(?) version of the arms the bar will extend inside the spring bundle and beyond the hoop. As in current arms the end of the bar will be bent to 90 degree angle and sunk into a small hole at the bottom of the groove in the cone. This will prevent any forward movement of the bar. The bar will also protrude clearly above the surface of the cone, so it contacts the field-frame bar first. I probably won't attach the bar and the hoop together, given that there haven't been any issues with a loose hoop so far. The design of the arms will be for all intents and purpose identical to what is shown in the cheiroballistra manuscripts and should work perfectly. The bowstring was also made much lighter. Previously I had used bowstrings weighing 12.94 and 9.61 grams. Today's bowstring weighed only 6.03 grams. All were made from nylon. I could probably make an even lighter bowstring, but I deem the current incarnation light enough for now. ---- So, what about the results after all these changes? The amount of washer rotation was about 180 degrees which seems less than what I've used in previous higher-power tests because: * It was possible to draw the weapon to -8.2 cm (of maximum draw) with fairly small effort * One arm had managed moved a bit forward in the bundle * Cords in one field-frame managed to slip over the hoop None of these were true with earlier tests. Interestingly there was even less noise than previously, which points at less energy being wasted. I definitely should start measuring the washer rotation from now on and - when possible - start chronographing a few shots with a bolt of known weight to ensure the energy levels match what I had used before. Anyways, here are the results of the latest shooting session: DRAW BOLT VELOCITY ENERGY SHOTS -21.7 cm 29.28 g 51.28 m/s 38.69 J 5 -14.9 cm 55.73 m/s 45.69 J 5 -8.2 cm 58.58 m/s 50.47 J 5 -21.7 cm 20.05 g 58.78 m/s 34.64 J 5 -14.9 cm 62.52 m/s 39.19 J 4 -8.2 cm 66.56 m/s 44.41 J 1 The 20 gram bolt had 88% of the energy of the 30 gram bolt. Assuming bolt energy drops linearly as bolt weight is reduced, a 10 gram bolt should have ~38.4 joules of energy. This would translate to a velocity of 87.59 m/s. The bolt energy will probably drop more sharply, but I'd still expect at least 80 m/s for a 10 gram bolts. It should be possible to add at least 50% more energy and still be able to cock the weapon with a stomach push alone. More energy, probably an additional 50% or more could be added by either grabbing and pulling the arms at the end of the draw, or by making the now loop-like handle a real handle that would be grabbed with both hands. This would ease the draw especially at the end, where the draw weight rises fairly sharply. Next in my production line are new arms with new bars and hoops. It is highly likely they will represent the peak of their development, and incidentally also look pretty much exactly the same as the arms in the cheiroballistra manuscripts. I will also ramp up the power levels, bind the halves of the spring bundle together with twine to prevent them slipping over the hoop, fix the tenon slippage issue and add a simple spring to make the handle auto-lock into the pin. Later I'll remake the entire dovetailed portion of the case using a less pronounced angle. Obviously a matching slider is also needed, and a composite construction with elm forming the rear part and birch forming the front part probably makes most sense. Elm is very hard to split, so it is optimal for attaching the handle, but using it for the entire slider would seem like a waste. At some point I will experiment with a real handle that can be pulled with both hands and which is also used to lock the slider. Category:Blog posts Category:Backup Category:Cheiroballistra Category:Lock Category:Arms Category:Case and slider Category:Little ladder Category:Little arch Category:Bowstring Category:Statistics